Speciation of Metals in Asphaltenes by High-Performance Thin-Layer

Jun 11, 2019 - (Figure S1) Processing data with FIJI; (Figure S2) calibration curve for each ... obtained by HPTLC-LA-ICP-MS after image processing (P...
0 downloads 0 Views 2MB Size
Article Cite This: Energy Fuels 2019, 33, 6060−6068

pubs.acs.org/EF

Speciation of Metals in Asphaltenes by High-Performance ThinLayer Chromatography and Laser Ablation Inductively Coupled Plasma-Mass Spectrometry Rémi Moulian,†,‡,§ Sara Gutierrez Sama,†,‡,§ Carole Garnier,‡ Sandra Mounicou,† Maxime Enrico,† Xavier Jaurand,∥ Ryszard Lobinski,†,§ Pierre Giusti,‡,§ Brice Bouyssiere,*,†,§ and Caroline Barrère-Mangote*,‡,§ Downloaded via UNIV OF BATH on August 20, 2019 at 08:06:17 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



CNRS/University of Pau/E2S, Institut des Sciences Analytiques et de Physico-Chimie pour l’Environnement et les Materiaux, UMR5254, Hélioparc- 2 Av Pr Angot, 64000 Pau, France ‡ TOTAL Refining and Chemicals, Total Research and Technologies, Gonfreville, BP 27, 76700 Harfleur, France § TOTAL RC, CNRS, Univ. Pau, Univ. Rouen, International Joint Laboratory − C2MC: Complex Matrices Molecular Characterization, TRTG, BP 27, 76700 Harfleur, France ∥ Centre Technologique des Microstructures, Université Claude Bernard Lyon1, 5 rue Raphael Dubois-Bâtiment Darwin B, F-69622 Villeurbanne, Cedex, France S Supporting Information *

ABSTRACT: Asphaltenes are considered to be the most problematic components of heavy oils because they can self-aggregate which leads to precipitation and causes various problems during oil recovery, transportation, and refining. The contribution of the porphyrins present in asphaltenes to the aggregation was previously studied by gel permeation chromatography inductively coupled plasma-mass spectrometry (GPC-ICP MS). The molecular weight of asphaltene aggregates was shown to be increased by free metal-containing porphyrins (corresponding to the lower molecular weight fraction) interacting with the aggregate’s surfaces by weak forces. The characterization of free porphyrins within the asphaltenes is therefore for the understanding of the mechanism of the aggregation, coprecipitation, and demetalation processes. Here, we developed a method for the separation of free porphyrins from asphaltenes on the basis of their polarity using high-performance thin-layer chromatography (HPTLC). This technique, using disposable plates, is particularly well suited for asphaltene analysis since it eliminates the risk of clogging typical of column chromatography. Cellulose plates were used in this study. The lower polarity of their hydroxyl groups limit the irreversible adsorption and improves the detection limit by the ICP-MS. Two well-separated peaks were obtained from purified asphaltene (Asphaltene 2017; Asphaltene Characterization Interlaboratory Study for PetroPhase 2017. In Proceedings of the 18th International Conference on Pretroleum Phase Behavior and Fouling, Le Havre, France, June 11−15, 2017; Total, the University of Pau, and the University of Rouen-Normandy: Le Havre, France, 2017.) and its corresponding whole crude oil and C5 and C7 fractions. The distribution of vanadium due to migration was determined by laser ablation (LA) ICP MS. The eluted fraction contained the free porphyrins, whereas the major fraction did not migrate and corresponded to trapped porphyrins. A comparison with the signal obtained by UV densitometry allowed the ratio between the inorganic and organic material to be measured.



INTRODUCTION The composition of asphaltenes, the most polar fraction of a petroleum matrix, remains largely unknown. Speight1 defined asphaltenes as the fraction that precipitates in the presence of alkanes but is soluble in hot toluene. The asphaltenes that precipitate in n-pentane, n-hexane, and n-heptane are referred to as C5, C6, and C7 asphaltenes, respectively. Asphaltenes contain a wide variety of molecules with different aromaticities, high polarities, and high contents of heteroatoms and metals (Ni, V···).2−4 These properties allow asphaltenes to selfaggregate in fairly concentrated solutions, leading to macroaggregation that results in precipitation in organic media.5−7 The nano and macroaggregation need to be understood to avoid problematic effects, such as pipe clogging. Also, nanoaggregation may be linked to the efficiency in catalytic processes via pore size. © 2019 American Chemical Society

Thin-layer chromatography (TLC), usually with silica plates, has been extensively8−18 used for the characterization of petroleum samples. For instance, Cebolla et al.8 used silica gel plates impregnated with berberine to separate SARA components (saturates, asphaltenes, resins, aromatics) of petroleum samples. Berberine was necessary to detect saturated compounds.8 Sharma et al.9 used TLC-FID (Iatroscan) to perform SARA separations with a standard deviation below 3.5%. Furthermore, Jarne et al.10 used modified silica gel plates impregnated with caffeine to separate petroleum products according to their number of rings. Mateos et al.11 used silica gel plates impregnated with a solution of Received: March 6, 2019 Revised: June 11, 2019 Published: June 11, 2019 6060

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels

characterization of the fractions by UV, mass spectrometry, and GPC. They obtained lower molecular weights and higher mobilities on the plates. Similar results were obtained by the same group for other samples.15 In this work, high-performance (HP) TLC is proposed for the speciation of asphaltenes, particularly for the separation and detection of “free porphyrins.” The porphyrins are detected by element specific techniques, such as laser ablation-inductively coupled plasma-mass spectrometry (LAICP-MS) and UV densitometry. Most of the vanadium and nickel ions in asphaltenes are part of porphyrin complexes (tetrapyrrolic complexes mainly of vanadyl (VO2+) and nickel(II) ions).24 Although porphyrins can be free in solution, they can also be trapped into nanoaggregates of asphaltenes, making them difficult to separate and analyze with UV−visible spectrophotometry.24,25 Improving the understanding the porphyrins’s role associated with asphaltene nanoaggregates by different types of interactions26,27 in the aggregation process is the main objective of this study. For this purpose, our recently developed method using GPC and HPTLC characterization is applied to different asphaltene samples obtained from the same crude oil with different alkanes, as well as to the Asphaltene 2017 reference sample.

coralyne cation, which is more sensitive than berberine for many compounds. The use of TLC with single-use plates for asphaltene characterization can avoid the damage to chromatography columns caused by irreversible adsorption of the polar compounds present onto the stationary phase and asphaltene precipitation. Recently, Chacon-Patino et al.12 and Giraldo-Davila et al.13 used silica gel plates to separate asphaltene samples. They used a mixture of DCM/MeOH and then toluene to separate asphaltenes in three fractions.12,13 This technique can also be used on the preparative scale to separate fractions that can then be scratched, extracted with solvents, and further analyzed by other techniques. 14 Vorapalawut et al.15 used TLC in conjunction with LA-ICPMS for SARA separation with silica-based plates for the speciation of sulfur, vanadium, and nickel from the SARA fractions of petroleum samples. Chirinos et al.16 also used this hyphenated technique for the rapid determination of the vanadium/nickel ratio in different samples of asphaltenes.16 Although silica-based plates are typically used for HPTLC for hydrocarbon-type analyses, these plates are affected by irreversible adsorption at the application point due to the high polarity of the silanol groups present, and this effect is most dramatic in asphaltene analysis. The use of less polar sorbents, such as cellulose, could allow a more robust speciation of asphaltene. The use of sorbents that are not based on silica could facilitate the detection of certain analyte molecules by eliminating the high silica signal derived from the plate in LA-ICP-MS. Several authors have used TLC for the separation of porphyrins and metalloporphyrins. In 1978, Hajibrahim et al.17 used preparative TLC for the separation of carotenoids and porphyrins for further analysis. In 1989, van Berkel et al.18 used liquid chromatography to obtain vanadyl porphyrin- and nickel porphyrin-enriched fractions from bitumen samples. Then, they used TLC with silica-based plates to separate the different porphyrins in each of the previous fraction. All the Ni porphyrins eluted with heptane/ DCM 3/2% v/v, while vanadyl porphyrins need a more polar solvent, such as 100% DCM. Vargas et al.19 used liquid chromatography to obtain fractions of crude oil with different mixtures of solvent ((1) hexane/DCM 9/1; (2) hexane/DCM 3/1; (3) hexane/DCM 1/1; (4) DCM/acetone 1/1; (5) DCM/methanol 4/1; (6) chloroform/methanol/water 65/25/ 4). In the second step, they used HPTLC by doing a SARA separation in order to determine what they had in each of the different fractions. They concluded that fractions 4, 5, and 6 are retained at the application point and correspond to asphaltene. At the same time, they used UV analysis to analyze each fraction and concluded that free vanadyl porphyrins are more present in fraction 3. This study aims to test for the first time the use of cellulosebased HPTLC plates for asphaltene speciation and to compare the results with those achieved with classic silica-based plates. The efficiency of the method for separating free porphyrins from aggregates is assessed by the analysis of samples previously separated by GPC.20−22 GPC is a widely used technique for the characterization of petroleum samples due to its ability to separate compounds by their hydrodynamic volume, molecular size, and/or aggregation state. This separation strategy uses a different principle to the polaritybased separation usually achieved on silica stationary phases. Deelchand et al.23 proposed an off-line hyphenation of TLC and GPC techniques involving a preparative TLC separation of petroleum residue and pitch on silica-based plates followed by



EXPERIMENTAL SECTION

Samples, Materials, and Reagents. Samples. Asphaltene 2017 (from a Middle East crude oil, with a vanadium content of 640 ppm) from the interlaboratory study of the PetroPhase congress, and its preparative GPC fractions,21 asphaltenes C5 and C7, were studied. The three samples (A2017, AC7, and AC5) came from the same crude oil. The difference between A2017 and AC7 is that A2017 is washed more after the precipitation (see the complete procedure https://petrophase2017.sciencesconf.org/resource/page/id/10). Asphaltene 2017 is obtained first by the ASTM D6560-12 method and then by a Soxhlet wash with heptane for 5 h. This step is repeated until the material extracted from asphaltene is below 0.2 mg/200 mL heptane. Two reference molecules (5,10,15,20-tetraphenyl-21H,23Hvanadyl porphyrin and 5,10,15,20-tetraphenyl-21H-23H-porphine nickel, purity 95%, Merck, Frankfurter Strasse 250, Darmstadt 64293, Germany) were used. Solvents. Tetrahydrofuran (THF, HPLC grade, without stabilizer, Lichrosolv, Merck, Frankfurter Strasse 250, Darmstadt 64293, Germany) was used as the solvent for sample preparation. To facilitate the release of porphyrins, the solutions of the whole asphaltenic samples were subjected to ultrasound for 1 h. Dichloromethane (DCM) and methanol (MeOH) (HPLC grade, Lichrosolv, Merck, Frankfurter Strasse 250, Darmstadt 64293, Germany) were used to develop the HPTLC plates. Toluene, heptane, and pentane were used to prepare Asphaltenes 2017, C7, and C5. HPTLC Plates. Classic HPTLC silica gel plates (thickness of 0.15− 0.2 mm, Merck, Frankfurter Strasse 250, Darmstadt 64293, Germany) and cellulose HPTLC glass plates (thickness of 0.06−0.11 mm, Merck, Frankfurter Strasse 250, Darmstadt 64293, Germany), with dimensions of 20 cm × 10 cm without fluorescent indicators were used. The cellulose stationary phase is thinner and contains smaller, more uniform particles than silica gel for faster separation, better separation efficiency, better resolution, and more sensitive detection.28 Instrumentation. HPTLC. The HPTLC procedure was carried out with the following instrumentation from Camag. The sample application and sample development were performed using an Automatic TLC Sampler 4 (ATS4) and an Automated Multiple Development 2 (AMD2), respectively. UV absorbance chromatograms were monitored at a wavelength of 280 nm with a D2&W lamp by a TLC scanner. UV spectra from 190 to 900 nm were obtained using the same device. Finally, a TLC visualizer was used to obtain a 6061

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels picture of the plate under 366 nm UV light after development. Peak area data were collected using WinCATS software. HPTLC plates were cleaned with DCM and THF and dried at 110 °C for 10 min prior to sample deposition. The samples were deposited 10 mm from the bottom of the plate. A high-strength mobile phase (DCM/MeOH, 99.5:0.5) was used to completely elute the free porphyrins, so they could be characterized separately from the rest of the material. The plate was developed in a single step until the solvent front reached 50 mm.29 The nickel and vanadyl porphyrins were calibrated by depositing six different amounts of the complexes, from 0.00537 to 1.08 μg for vanadyl porphyrins and from 0.0054 to 1.08 μg for nickel porphyrins. Furthermore, three different amounts of Asphaltene 2017, C5, and C7 (1, 3, and 5 μg) were deposited. All the details are available in Table S2 of the Supporting Information. The calibration standards and the samples were deposited on three cellulose plates for repeatability and on a silica plate for comparison. The repeatability of the method was calculated, and the differences between silica and cellulose were determined. Additionally, Asphaltene 2017 (1, 2.5, and 5 μg) and its GPC fractions (1, 2.5, and 5 μg) were deposited on another cellulose plate. LA-ICP MS. Element specific analysis of the developed HPTLC plate was carried out by LA-ICP MS using a New Wave Research UP213 Nd:YAG laser (ESI, Fremont, CA) coupled with ICP MS (7700 series, Agilent, Santa Clara, CA) under the operating conditions given in Table 1. The instrument was equipped with a Fassel-type quartz

The preparative GPC system was calibrated with nine polystyrene standards with Mw values ranging from 630 to 2350000 Da using a refractive index detector. The equation of the mass calibration curve was log (Mw) = −1.10−5x3+ 0.0048x2 −0.7188x + 40.614 (R > 0.9988), where x is the elution volume. One milliliter of the asphaltene solution diluted 100 times in THF (Sigma-Aldrich) was injected into the column by means of an injection loop connected to an external 6-way valve. Xylene was used as the mobile phase instead of THF, and the eluent was delivered at a flow rate of 3 mL/min by a Dionex high-performance liquid chromatography (HPLC) system with an UltiMate 3000 microflow pump. Fractions of different molecular weights were collected at the outlet of the preparative columns, and the fractions contained the highmolecular-weight aggregates (HMW), medium-molecular-weight (MMW) compounds, low-molecular-weight (LMW) compounds, and the tailing fraction of compounds retained on the stationary phase. The peak corresponding to the porphyrinic solution was observed in the LMW fraction; thus, the free porphyrins are expected to be present in this fraction.



RESULTS AND DISCUSSION Reference Metalloporphyrins. As shown by Chirinos et al.,16 in order to obtain measurable LA-ICP-MS signals using classic Si-based plates, several tens of micrograms of crude oil are typically deposited. This amount of sample does not allow good quality migrations in HPTLC because of the thin layer of sorbent (approximately 100 μm vs 250 μm for classic TLC15,16). Cellulose plates represent an interesting alternative for achieving suitable migration. Indeed, unlike silica, cellulose is a polar sorbent with alcohol and ether functional groups that is suitable for adsorption chromatography. Moreover, the use of cellulose plates improves the detection in ICP MS by removing the intense silica interferences and improving the plasma stability. Figure 1 compares silica- and cellulose-based HPTLC plates with 1 μg of each reference vanadyl and nickel porphyrin deposited. The detections were carried out by LAICP MS. The 51V signal intensities from the two kinds of HPTLC plates differed by a factor of 10 (Figure 1a). As

Table 1. LA-ICP MS Operating Parameters Agilent 7700 Series ICP MS plasma Ar gas flow rate (L min−1) 15 auxiliary Ar gas flow rate (L min−1) 1 13 measured isotope C, 32S, 51V, 58Ni, 60Ni analysis time (s) 506 New Wave Research UP-213 LA wavelength (nm) pulse energy (mJ) fluence (J cm−2) spot size (μm) repetition rate (Hz) scan speed (μms−1) carrier He gas flow rate (L min−1)

213 0.45 (50%) 3.5 200 20 100 0.5

torch shielded with a grounded Pt electrode and a quartz bonnet. A standard quartz injector (1.75 mm i.d.) was used. An ablation chamber was installed on a three-axis translation stage and coupled to the ICP torch using a 60 cm Tygon tube (5.0 mm i.d.). The ablated material was swept by the carrier gas (helium) and mixed with the spectrometer makeup gas (argon) prior to introduction into the plasma. Ni sampler (orifice diameter of 1.1 mm) and skimmer (orifice diameter of 0.8 mm) cones were used.15 The laser beam followed a linear trajectory throughout the migration (see Table S1 for repeatability on one sample). The distribution of 51V compounds was controlled. To fit within the ablation cell, plates were cut to dimensions of 9 × 10 cm. FIJI30,31 image processing software was used to precisely determine the area of each spot of sample on the solvent front and normalize all the results obtained by UV and LAICP MS. Preparative GPC. Fractionation was carried out following the method of Putman et al.21 Briefly, we used three Shodex (Showa Denko, Japan) polystyrene−divinylbenzene (PS/DVB) preparative gel permeation columns (20 mm i.d., 300 mm length), i.e., KF2001 (particle size: 6 μm, exclusion limit: 1500 Da), KF2002.5 (particle size: 6 μm, exclusion limit: 20000 Da), and KF2004 (particle size: 7 μm, exclusion limit: 400000 Da) connected in series with a KF-LG precolumn for protection. The column temperature was not controlled.

Figure 1. (a) 51V signal and (b) 58Ni signal from LA-ICP-MS of a one-step (40 mm) development with silica (red line) and cellulose (black line) HPTLC plates with DCM/MeOH 99.5/0.5% v/v of 1 μg of the reference porphyrin mixture. 6062

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels

Figure 2. Densitometry UV spectrum of the (black line) eluted and (red line) noneluted fractions of the reference porphyrin mixture on cellulose (a) and silica (b) HPTLC plates. The UV signal at 280 nm (red line) and the 51V signal (black line) from LA-ICP-MS of the reference porphyrin mixture on cellulose (c) and silica (d) HPTLC plates.

two plates (Figure 2). Even when the UV spectra were recorded from the solid HPTLC plate rather than in solution, the expected absorption bands for the porphyrins were observed in the spectra of the eluted fractions on the plates with a large Soret band at approximately 425 nm (gray band), a α band at approximately 545 nm, and an β band at approximately 575 nm (Figure 2 (a, b vertical black lines)). The wavelengths of the maximum intensities of these bands were close to those reported by Freeman et al.33 for benzo vanadyl porphyrins. The spectra obtained from the noneluted fractions on the plates are shown in Figure 2 (a, b red lines). The spectrum obtained from the cellulose plate was noisier, and no characteristic absorption bands indicating the presence of the vanadium porphyrin complexes at the application point were observed. In contrast to that, the spectrum obtained at the application point from the silica-based plate was similar to the spectra obtained for the porphyrins at the elution front, suggesting that a portion of the analyte was adsorbed at the application point. Futhermore, when a cellulose plate was used (Figure 2 (c)), the UV signal (red line) at the deposition point was only 1% of the total signal. Note that in the case of the LAICP-MS 51V signal (black line), the signal at the deposition point was 6% of the total. This would mean that there is 6% of the vanadium at the application point but that it is not inside a

expected, vanadyl porphyrins quantitatively elute at the solvent front on the cellulose plate (black line). In comparison, only 85% of these species migrate with the solvent front (red line) on silica-based plates, indicating that a significant fraction remained at the application point. Similar results were obtained for nickel porphyrins (Figure 1b). Our observations confirm that cellulose interacts less with the porphyrin samples than does the silica gel. Moreover, the 58Ni LA-ICP MS signal is half as intense as the vanadium signal (V has a lower ionization potential (6,7) than Ni (7,6) and thus has a better detection limit in ICP MS). From the chromatographic point of view, the separation of the components is likely to be the result of their different solubilities in the solvent. Our purpose is to understand what migrates or does not migrate with a mixture like DCM/MeOH. This unexpected retention of porphyrins at the application point could be due to irreversible adsorption, as previously observed in our laboratory for polar compounds on silica plates. Several authors reported the degradation of metalloporphyrins on silica-based TLC plates, but degradation preferentially occurred for Ni porphyrins, as shown by Ali et al.32 and van Berkel et al.18 To better understand this phenomenon, UV−vis densitometry spectra were recorded directly at the application point and at the solvent front on the 6063

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels

R2 = 0.994 and F = 1325. This calibration curve was prepared for a single cellulose plate. The others calibration curves on plate 2 and 3 are given in Figure S2. The results of the interplate repeatability are given in Table 2. These results indicate that the relative standard deviation from one plate to another could be up to 25%, probably resulting from differences in migration and plate quality. For this reason, calibration points were added to each plate (A, B, and C) and were used to calibrate the quantification data. GPC Fractions of Asphaltene 2017. Our previous work20 using GPC showed that heavy petroleum samples could be separated into four fractions, referred to as the highmolecular-weight (HMW), medium-molecular-weight (MMW), low-molecular-weight (LMW), and tailing fractions for analytes having interactions with the stationary phase (unexpected separation mechanism in GPC), as proposed in Figure 4. Note that by GPC-ICP-MS, our reference vanadyl porphyrins each lead to a sharp peak in the LMW area,20,22 as shown in Figure 4.

porphyrin. Opposite that, on silica, as shown in Figure 2 (d), at the application point, 15% of the vanadium and 18% of the UV signal have been detected. This would mean that on silica, more vanadium is retain and at least a portion is vanadyl porphyrin. Vanadium calibration when the cellulose plates were used showed that the reference porphyrin moved with the front solvent (Figure 3). The spead of the peaks at the solvent front

Figure 3. (a) Picture at 280 nm of the cellulose plate after separation of the vanadyl porphyrins. (b) Six-point calibration curve obtained with vanadyl porphyrin by LA-ICP-MS after image processing. Figure 4. Chromatogram of vanadium in Asphaltene 2017 from GPCICP-MS (black line) and vanadyl porphyrin (dotted line). Fraction limits were set at 39−49 min for HMW, 49−60 min for MMW, 60− 70 min for LMW, and 71−91 min for tailing based on the trimodal distribution seen in chromatograms of V-containing aggregates.

and the absence of peaks in the middle of the plate can be explained by a “solubility mode” separation rather than a classic chromatographic separation. Indeed we could suppose that, on the cellulose phase with very polar solvent as DCM/ MeOH, all molecules that have been dissolved are eluted without any retention. This plate was analyzed by LA-ICP MS, but the size of the eluted spot from the different migration trials and on different plates differed because the spot spread out. To solve this problem, we used image processing software (FIJI) to determine the area of each spot of the eluted and noneluted fractions. The laser-ablated area together with ICP MS signal data allowed the correction of the data (Figure S1). FIJI was used for the treatment of all the plates. The calibration curve obtained for vanadyl porphyrin (shown in Figure 3b) was y = 2 × 108x (±5.5 × 106) with

A 1 mL aliquot of Asphaltene 2017 diluted in THF was fractionated according to this method. The obtained chromatogram is shown in Figure 4, and the corresponding mass balance is given in Table 3. There is no correlation between the mass of the fraction and quantity of vanadium, and most of the metal species and total material are part of the HMW fraction. However, the percent of vanadium in the LMW fraction is higher than the percent of total material in this fraction.

Table 2. Calibration with Vanadyl Porphyrins vanadium (μg)

area A

area B

area C

average

standard deviation (%)

0.00537 0.0108 0.0537 0.108 0.537 1.08

524304 1456995 7797726 13860441 118697108 209863919

676552 1404246 7976937 16379744 162436773 231665273

706549 896046 6050314 12440338 108242613 169634752

635802 1252429 7274992 14226841 129792165 203721315

15 24 14 14 22 15

6064

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels Table 3. Mass Balance of Asphaltene 2017 and 51V Repartition in the Collected GPC Fractions Obtained by GPC-ICP-MS

mass of Asphaltene 2017 (mg) mass of Asphaltene 2017 (%) vanadium by GPC-ICP-MS (%)

Injected Quantity

HMW

MMW

LMW

Tailing

9.0

4.8

2.3

0.9

0.2

8.2

53.3

25.6

10.0

2.2

91.1

52.3

26.9

15.0

5.8

and that these porphyrins are more polar (polar side chains) or they can reaggregate. However, Deelchand et al.23 observed the same trend; higher molecular weight fractions showed lower mobilities on the HPTLC plate. The same tendency is observed for deposits of 2.5 and 5 μg (Table S3). We can also follow the ratio between LA and UV (Table 4). For the MMW and LMW fractions of Asphaltene 2017, the ratio of the eluted part of LA/UV is between 600 and 800. However, the ratio is approximately 300 for the HMW fraction. Thus, the eluted MMW and LMW fractions are more concentrated in vanadium than the HMW is. An UV−vis spectrum for each peak was recorded to obtain structural information on the molecules present, and the spectra are given Figure 6. The spectrum obtained for the noneluted part, drawn in red, is similar for each fraction. This spectrum, with a maximum intensity at approximately 240 nm that decreases continuously until 800 nm, is characteristic of highly polyaromatic compounds. These spectra are similar to those obtained by Bonoldi et al.34 for an asphaltene diluted in THF. However, unlike in the work of Bonoldi et al.,34 no peaks were observed at approximately 400 nm, which is the wavelength of the porphyrin Soret band. On the other hand, large peaks were observed in LA-ICP-MS at the application point, indicating the presence of vanadium. As proposed by Evdokimov et al.,25 this could be due to the occlusion of porphyrins inside the aggregates, which would prevent the detection of their characteristic UV−vis absorption bands. Furthermore, the spectra obtained from the eluted parts (black line) differed depending on the fraction. For the HMW fraction (Figure 6), the signal between 300 and 400 nm is weaker. For the MMW, LMW, and T fractions, we can see the Soret band of the porphyrins at 420 nm, especially for the LMW fraction (represented by a star). This result suggests that separation by HPTLC allows the isolation of the free porphyrins contained in the MMW, LMW, and T fractions from the rest of the material. Asphaltenes 2017, C7, and C5. The distribution of the 51 V species in Asphaltene 2017, Asphaltene C7, and Asphaltene C5 was controlled using the same methodology, and the distributions are given in Figure 7 and Table 5. As shown in Figure 7, again only two peaks were observed. The fraction of the asphaltene with the greatest affinity for the plate has a higher vanadium content than the eluted fraction (Table 5), meaning that the largest portion of the metal would be in the aggregates. Table 5 shows that the ratios between LA/UV for the application point of asphaltene varied. For Asphaltene C7 and 2017, the ratio is close to 1800. This would mean that the molecules removed during the washing do not stay at the application point or are not aromatics (as UV mostly detects aromatic compounds). Therefore, for Asphaltene C5, the ratio is close to 1300, this would mean that in Asphaltene C5 there is more organic aromatic matter than in the others asphaltenes. For the eluted part, the ratio between LA/UV for Asphaltene 2017 is close to 900 instead of 500 for C5 and C7. This would mean that precipitate asphaltene with C7 and wash it allow a better migration of metallic compounds. The calibration curves shown in Table 2 were used to calculate the concentrations of vanadium in Asphaltene 2017, C5, and C7 (Table S4). The average concentration on the three different plates with three different depositions was approximately 642 ppm, which is in good agreement with the concentration obtained by the ATSM method (640 ppm). Thus, HPTLC-LA-ICP-MS with a cellulose plate allowed us to

Recovery

100

The different GPC fractions were deposited on HPTLC cellulose plates and developed using the same procedure as was used for the porphyrins. The plates were analyzed by LA-ICP MS and by UV densitometry at 280 nm. A picture of the plate and the LA-ICP MS chromatograms are given in Figure 5, and

Figure 5. (a) Picture at 280 nm of the cellulose plate after migration. (b) LA-ICP-MS chromatograms (black line) and UV spectra (red line) of the one-step development on cellulose HPTLC plates with DCM/MeOH of 1 μg of the preparative GPC fractions of Asphaltene 2017.

the various measured areas are given in Table 4. Figure 5 shows that for all GPC fractions, a portion of the sample remains at the application point and a portion migrates with the solvent front. Surprisingly, even for the LMW fraction, which is considered the free porphyrin part of the sample, 65% of the vanadium remains at the application point, as indicated in Table 4. This result means that the LMW fraction is not only made of porphyrins similar to our porphyrin standards 6065

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels

Table 4. Area and Percentages of Noneluted and Eluted Material from 1 μg of Each Fraction for UV and 51V by LA-ICP-MS After Image Processing area sample

noneluted

eluted

percentage total

noneluted

eluted

LA/UV noneluted

LA/UV eluted

2530 601 762 623 508 2722018

798 274 866 639 307 531301

UV A 2017 HMW MMW LMW T porphyrins

1311 1152 2113 1446 1654 5

588 232 704 754 966 395

2556 1354 2814 2200 2540 400

77 83 75 66 63 1

23 17 25 34 37 99

A 2017 HMW MMW LMW T porphyrins

3317643 692888 1609144 900274 839530 13610088

469063 63514 609555 482428 297053 209863919

LA-ICP-MS 3786707 756402 2218699 1382702 1136583 223474007

88 92 73 65 74 6

12 8 27 35 26 94

Figure 7. LA-ICP-MS (black line) and UV (red line) chromatograms of the one-step development on cellulose HPTLC plates with DCM/ MeOH of 1 μg of Asphaltene 2017 (a), Asphaltene C7 (b), and Asphaltene C5 (c).

Figure 6. UV spectra of 1 μg of each preparative GPC fraction, i.e., HMW, MMW, LMW and T, of Asphaltene 2017 and the total Asphaltene 2017 at the application point (red line) and at the solvent front (black line). Dotted lines represent the vanadyl porphyrin. Vertical lines represent a change in the lamp setting on the device. The Soret band corresponding to free porphyrins is indicated with a star. .



CONCLUSIONS In this work, instead of classic Si-based sorbents, cellulose plates were used for HPTLC analysis of petroleum samples. Cellulose leads to 10-fold more intense signals with less noise than silica gel when the plates are analyzed by LA-ICP MS and allows the quantification of vanadium forms in asphaltene samples. Two well-separated peaks were obtained for purified asphaltene (Asphaltene 2017)35 and its corresponding fractions. The eluted fraction was found to contain free porphyrins, whereas the fraction that remained at the application point (the major fraction) corresponded to trapped or highly polar porphyrins and was the most difficult part to

calculate the concentration of vanadium in the asphaltene sample. According to the results, the concentration in Asphaltene C7 has an average of 611 ppm vanadium is close to that in Asphaltene 2017, which has an average of 642 ppm of vanadium. However, Asphaltene C5 is less concentrated (464 ppm), meaning that during the precipitation of asphaltene, heptane promotes more coprecipitation of metals than nonmetallic compounds. 6066

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels

Table 5. Average of the Areas and Percentages of Noneluted and Eluted Material from 5 μg of Each Asphaltene Sample on Three Different Plates for UV and 51V by LA-ICP-MS After Image Processing area sample

noneluted

eluted

percentage total

A 2017 AC 7 AC 5 porphyrins

3926 3875 3644 5

954 935 1175 395

4880 4810 4819 400

A 2017 AC 7 AC 5 porphyrins

7307202 6878423 4932036 13610088

918289 433202 581532 209863919

8225490 7311625 5513568 223474007

noneluted UV 80 81 76 1 LA-ICP-MS 89 94 89 6

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.9b00676.



(Figure S1) Processing data with FIJI; (Figure S2) calibration curve for each plate A, B, and C; (Table S1) intensity of three replicates on the same spot (AC5 3 μg) by LA-ICP-MS; (Table S2) mass deposit for asphaltene and porphyrins samples; (Table S3) results obtained for 1, 2.5, and 5 μg of deposit of A2017 and fractions by UV and LA-ICP-MS; and (Table S4) concentration obtained by HPTLC-LA-ICP-MS after image processing (PDF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Tel.: +33 (0) 235 551 102. *E-mail: brice.bouyssiè[email protected]. Tel.: +33 (0) 559 407 752. ORCID

Pierre Giusti: 0000-0002-9569-3158 Brice Bouyssiere: 0000-0001-5878-6067 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the Conseil Régional d’Aquitaine (20071303002PFM) and FEDER (31486/08011464). The authors thank TOTAL for supplying the oil samples.



ratio LA/UV roneluted

ratio LA/UV eluted

20 19 24 99 11 6 11 94

1861 1775 1353 2722018

963 463 495 531301

(2) Strausz, O. P.; Peng, P. A.; Murgich, J. About the colloidal nature of asphaltenes and the MW of covalent monomeric units. Energy Fuels 2002, 16, 809−822. (3) Leyva, C.; Ancheyta, J.; Berrueco, C.; Millán, M. Chemical characterization of asphaltenes from various crude oils. Fuel Process. Technol. 2013, 106, 734−738. (4) Pereira, T. M. C.; Vanini, G.; Oliveira, E. C. S.; Cardoso, F. M. R.; Fleming, F. P.; Neto, A. C.; Lacerda, V.; Castro, E. V. R.; Vaz, B. G.; Romão, W. An evaluation of the aromaticity of asphaltenes using atmospheric pressure photoionization Fourier transform ion cyclotron resonance mass spectrometry − APPI(±)FT-ICR MS. Fuel 2014, 118, 348−357. (5) McKenna, A. M.; Marshall, A. G.; Rodgers, R. P. Heavy petroleum composition. 4. Asphaltene compositional space. Energy Fuels 2013, 27, 1257−1267. (6) Barrera, D. M.; Ortiz, D. P.; Yarranton, H. W. Molecular weight and density distributions of asphaltenes from crude oils. Energy Fuels 2013, 27, 2474−2487. (7) Yarranton, H. W.; Ortiz, D. P.; Barrera, D. M.; Baydak, E. N.; Barré, L.; Frot, D.; Eyssautier, J.; Zeng, H.; Xu, Z.; Dechaine, G.; Becerra, M.; Shaw, J. M.; McKenna, A. M.; Mapolelo, M. M.; Bohne, C.; Yang, Z.; Oake, J. On the size distribution of self-associated asphaltenes. Energy Fuels 2013, 27, 5083−5106. (8) Cebolla, V. L.; Membrado, L.; Domingo, M. P.; Henrion, P.; Garriga, R.; González, P.; Cossío, F. P.; Arrieta, A.; Vela, J. Quantitative applications of fluorescence and ultraviolet scanning densitometry for compositional analysis of petroleum products in thin-layer chromatography. J. Chromatogr. Sci. 1999, 37, 219−226. (9) Sharma, B. K.; Sarowha, S. L. S.; Bhagat, S. D.; Tiwari, R. K.; Gupta, S. K.; Venkataramani, P. S. Hydrocarbon group type analysis of petroleum heavy fractions using the TLC-FID technique. Fresenius' J. Anal. Chem. 1998, 360, 539−544. (10) Jarne, C.; Cebolla, V. L.; Membrado, L.; Le Mapihan, K.; Giusti, P. High-performance thin-layer chromatography using automated multiple development for the separation of heavy petroleum products according to their number of aromatic rings. Energy Fuels 2011, 25, 4586−4594. (11) Mateos, E.; Cebolla, V. L.; Membrado, L.; Vela, J.; Galvez, E. M.; Matt, M.; Cossio, F. P. Coralyne cation, a fluorescent probe for general detection in planar chromatography. J. Chromatogr. A 2007, 1146, 251−257. (12) Chacón-Patiño, M. L.; Blanco-Tirado, C.; Orrego-Ruiz, J. A.; Gómez-Escudero, A.; Combariza, M. Y. High resolution mass spectrometric view of asphaltene−SiO2 interactions. Energy Fuels 2015, 29, 1323−1331. (13) Giraldo-Dávila, D.; Chacón-Patiño, M. L.; McKenna, A. M.; Blanco-Tirado, C.; Combariza, M. Y. Correlations between molecular composition and adsorption, aggregation, and emulsifying behaviors of petrophase 2017 asphaltenes and their thin-layer chromatography fractions. Energy Fuels 2018, 32, 2769−2780. (14) Li, W.; Morgan, T. J.; Herod, A. A.; Kandiyoti, R. Thin-layer chromatography of pitch and a petroleum vacuum residue. Relation

remove. These results suggest that is separation could be used to pilot efficient processes using a hydrodemetalation step. Moreover, the differences between Asphaltene 2017, C7, and C5 can be determined by comparing their UV and LA-ICP MS data as a function of the nature of the asphaltene. The concentration in vanadium is higher in asphaltene precipitated with heptane, and there is less organic matter in Asphaltene 2017.



eluted

REFERENCES

(1) Speight, J. G. Petroleum asphaltenes - Part 1: asphaltenes, resins and the structure of petroleum. Oil Gas Sci. Technol. 2004, 59, 467− 477. 6067

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068

Article

Energy & Fuels between mobility and molecular size shown by size-exclusion chromatography. J. Chromatogr. A 2004, 1024, 227−243. (15) Vorapalawut, N.; Martinez Labrador, M.; Pohl, P.; Caetano, M.; Chirinos, J.; Arnaudguilhem, C.; Bouyssiere, B.; Shiowatana, J.; Lobinski, R. Application of TLC and LA ICP SF MS for speciation of S, Ni and V in petroleum samples. Talanta 2012, 97, 574−578. (16) Chirinos, J.; Oropeza, D.; González, J.; Ranaudo, M.; Russo, R. E. Determination of Vanadium/Nickel proportionality in the asphaltene fraction of crude oil using thin-layer chromatography with femtosecond laser ablation−inductively coupled plasma−mass spectrometry. Energy Fuels 2013, 27, 2431−2436. (17) Hajibrahim, S. K.; Tibbetts, P. J. C.; Watts, C. D.; Maxwell, J. R.; Eglinton, G.; Colin, H.; Guiochon, G. Analysis of carotenoid and porphyrin pigments of geochemical interest by high-performance liquid chromatography. Anal. Chem. 1978, 50, 549−553. (18) Van Berkel, G. J.; Quirke, J. M. E.; Filby, R. H. The Henryville Bed of the New Albany shaleI. Preliminary characterization of the Nickel and Vanadyl porphyrins in the bitumen. Org. Geochem. 1989, 14, 119−128. (19) Vargas, V.; Castillo, J.; Ocampo Torres, R.; Bouyssiere, B.; Lienemann, C.-P. Development of a chromatographic methodology for the separation and quantification of V, Ni and S compounds in petroleum products. Fuel Process. Technol. 2017, 162, 37−44. (20) Gutierrez Sama, S.; Desprez, A.; Krier, G.; Lienemann, C.-P.; Barbier, J.; Lobinski, R.; Barrere-Mangote, C.; Giusti, P.; Bouyssiere, B. Study of the aggregation of metal complexes with asphaltenes using gel permeation chromatography inductively coupled plasma highresolution mass spectrometry. Energy Fuels 2016, 30, 6907−6912. (21) Putman, J. C.; Gutiérrez Sama, S.; Barrère-Mangote, C.; Rodgers, R. P.; Lobinski, R.; Marshall, A. G.; Bouyssière, B.; Giusti, P. Analysis of petroleum products by gel permeation chromatography coupled online with inductively coupled plasma mass spectrometry and offline with fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels 2018, 32, 12198. (22) Desprez, A. Caractérisation moléculaire et élémentaire des produits pétroliers lourds. Ph.D. Thesis, University of Pau, Pau, France, 2014. (23) Deelchand, J. P.; Naqvi, Z.; Dubau, C.; Shearman, J.; Lazaro, M. J.; Herod, A. A.; Read, H.; Kandiyoti, R. Planar chromatographic separation of petroleum residues and coal-derived liquids. J. Chromatogr. A 1999, 830, 397−414. (24) Zhao, X.; Xu, C.; Shi, Q. Porphyrins in heavy petroleums: a review. Structure and Modeling of Complex Petroleum Mixtures 2016, 39−70. (25) Evdokimov, I. N.; Fesan, A. A.; Losev, A. P. Occlusion of foreign molecules in primary asphaltene aggregates from near-UV− visible absorption studies. Energy Fuels 2017, 31, 1370−1375. (26) Caumette, G.; Lienemann, C.-P.; Merdrignac, I.; Bouyssiere, B.; Lobinski, R. Element speciation analysis of petroleum and related materials. J. Anal. At. Spectrom. 2009, 24, 263−276. (27) Barrow, M. P.; McDonnell, L. A.; Feng, X.; Walker, J.; Derrick, P. J. Determination of the nature of naphthenic acids present in crude oils using nanospray fourier transform ion cyclotron resonance mass spectrometry: the continued battle against corrosion. Anal. Chem. 2003, 75, 860−866. (28) Striegel, M. F.; Hill, J. Thin-layer Chromatography for Binding Media Analysis. Getty Conservation Institute: Los Angeles, CA, 1996. (29) Poole, C. F.; Dias, N. C. Practitioner’s guide to method development in thin-layer chromatography. J. Chromatogr. A 2000, 892, 123−142. (30) Rueden, C. T.; Schindelin, J.; Hiner, M. C.; DeZonia, B. E.; Walter, A. E.; Arena, E. T.; Eliceiri, K. W. ImageJ2: ImageJ for the next generation of scientific image data. BMC Bioinformatics 2017, 18 (1), 1. (31) Schindelin, J.; Arganda-Carreras, I.; Frise, E.; Kaynig, V.; Longair, M.; Pietzsch, T.; Preibisch, S.; Rueden, C.; Saalfeld, S.; Schmid, B.; Tinevez, J.-Y.; White, D. J.; Hartenstein, V.; Eliceiri, K.; Tomancak, P.; Cardona, A.; et al. Fiji: An open source platform for biological image analysis. Nat. Methods 2012, 9 (7), 676−682.

(32) Ali, M. F.; Perzanowski, H.; Bukhari, A.; Al-Haji, A. A. Nickel and vanadyl porphyrins in Saudi Arabian crude oils. Energy Fuels 1993, 7, 179−184. (33) Freeman, D. H.; Saint Martin, D. C.; Boreham, C. J. Identification of metalloporphyrins by third-derivative UV/VIS diode array spectroscopy. Energy Fuels 1993, 7, 194−199. (34) Bonoldi, L.; Flego, C.; Galasso, L. Beyond the average molecule description of asphaltenes: hyphenated gel permeation chromatography and spectroscopic analyses. Energy Fuels 2016, 30, 3630−3636. (35) PetroPhase. Asphaltene Characterization Interlaboratory Study for PetroPhase 2017. In Proceedings of the 18th International Conference on Pretroleum Phase Behavior and Fouling, Le Havre, France, June 11−15, 2017; Total, the University of Pau, and the University of Rouen-Normandy: Le Havre, France, 2017.

6068

DOI: 10.1021/acs.energyfuels.9b00676 Energy Fuels 2019, 33, 6060−6068